On Tue, May 29, 2012 at 7:14 PM, Matthew Wild <mwi...@gmail.com> wrote: > On 29 May 2012 18:46, Peter Saint-Andre <stpe...@stpeter.im> wrote: >> On 5/29/12 11:39 AM, Kevin Smith wrote: >>> On Sat, May 26, 2012 at 12:21 PM, Theo Cushion <t...@jivatechnology.com> >>> wrote: >>>> Dan, >>>> >>>> This is something we are interested in doing as well. Our use case is that >>>> people could be involved in multiple discussions, but only actually want >>>> to join a room if something interesting is going on (for example, missed >>>> messages or a user joining the room). The Pubsub mechanism would provide a >>>> great choice for this as it means the node can be configured so that the >>>> client does not have to request any of this information, or join a room in >>>> order to receive updates. >>> >>> The most elegant way of doing this would be to have a pubsub service >>> running on the MUC JID, just as PEP is a pubsub service running on a >>> user's JID. >> >> Right, we've traditionally called this "MEP" (MUC Eventing Protocol) but >> we've never defined the feature more fully than giving it a name. > > I did some experiments around this some time back: > http://blog.prosody.im/multi-user-chat-gets-rich/ > > However there are various approaches to tackling the problem, and I > gather that many people don't think mine was the best. We had a group > discussion on alternatives in Brussels in 2011, but again the > conclusion was that all approaches had drawbacks, some being > incredibly complicated - and so no consensus was reached.
Wasn't your post about a different problem? I thought it was about users publishing their own data (PEP through MUC) rather than having data attached to the room. /K _______________________________________________ JDev mailing list Info: http://mail.jabber.org/mailman/listinfo/jdev Unsubscribe: jdev-unsubscr...@jabber.org _______________________________________________